Gut microbiology of UK care home residents: a cross-sectional analysis from a randomised controlled trial

Objective: To describe the prevalence of potentially clinically relevant gut pathogens and associations with the carriage of resistant organisms in UK care home residents. Methods: Stool samples were collected pre-randomisation from care home residents participating in a randomised placebo-controlled trial. Cultivable clinically relevant bacteria were analysed. Antimicrobial susceptibility testing was performed by agar dilution (amoxicillin, co-amoxiclav, gentamicin, trimetho-prim, nitrofurantoin, and cipro ﬂ oxacin). We also aimed to detect resistance to third-generation cephalosporins, carbapenems, and vancomycin. Results: Stool samples were available for 159/310 residents participating in the trial (51%) from 23 care homes between 2016 and 2018. In total, 402 bacterial isolates were cultured from 158 stool samples and 29 different species were cultured. The ﬁ ve most common species were Escherichia coli (155/158, 98%), Pseudomonas aeruginosa (40/158, 25%), Enterococcus faecalis (35/158, 22%), Enterococcus faecium (30/158, 19%), and Proteus mirabilis (25/158,16%). Enterobacterales isolates were cultured from 157 samples (99%), and resistance to at least one of the tested antimicrobials was found in 119 of these (76%). There were high levels of variation in outcomes by care home. Discussion: We demonstrated that care home residents harbour signi ﬁ cant levels of antimicrobial-resistant organisms in their stool. This work emphasises the importance of both enhanced infection control practices and antimicrobial stewardship programmes to support the appropriate use of antimicrobials in this setting. David Gillespie, Clin Microbiol Infect 2023; ▪ :1 © 2023 The Author(s). Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases. This is an open access article under the CC BY license (http://creativecommons. org/licenses/by/4.0/).


Introduction
Care homes are enclosed environments of mostly frail older adult residents who are more likely to require living assistance, antibiotic treatment and invasive devices such as catheters [1]. Residents are at higher risk from infections, which can partially be attributed to infection control challenges such as increased risk of infection through diminishing immunity with age, increased proximity in sharing of living space, objects and bathroom facilities, low implementation of infection control procedures and increased antibiotic prescribing [1,2].
Care homes have been recognised as an important reservoir for multi-drug resistant organisms [3e6]. Infections are more likely to be caused by antimicrobial-resistant (AMR) bacteria [2,6e8], with urinary tract infections (UTIs) and respiratory tract infections being the most common healthcare-associated infection both in the hospital setting and in long-term care facilities (LTCFs) [9]. A large population study conducted in England demonstrated that residents in long-term care facilities are more than twice as likely as community-dwelling adults of similar age to present with a laboratory-confirmed Escherichia coli or Klebsiella UTI and four times more likely for the UTI to be caused by resistant bacteria [9].
Recurrent UTI infection has been linked with extended-spectrum (ES) cephalosporin resistance in Enterobacterales, which is a concern both in long-term care facilities and in the acute-care setting [10]. Many of the bacteria resistant to ES cephalosporins are extended spectrum beta-lactamase (ESBL)-producing organisms that are challenging to treat, with ESBL-producing organisms under surveillance by the World Health Organisation (WHO) as a key indicator of AMR evolution [11 ].
Furthermore, individuals living together have more similar gut bacterial communities than individuals living in other households [7]. Exploring the gut microbiological make-up in residents of care homes may lead to a better understanding of infection and AMR within this setting [8]. The high presence of multi-drug resistant bacteria in the gut, acting as a reservoir for potential future infections, may affect treatment options for this patient group.
Given the interplay between the microbiome and immunity and the additional risk factors associated with care homes and the elderly, interventions aiming to enhance the microbiome constitution may be one strategy to address AMR in this population. As AMR is not currently routinely assessed within this setting in the UK, or even well described, additional research within this area is warranted. In this study, we aimed to categorise and describe the clinically relevant potential pathogens and their associated antibiotic susceptibilities in the gut of UK care home residents, exploring variation between individuals and care homes. The study concentrated on relevant pathogens which are a frequent cause of infection in the elderly and is in contrast to microbiome research, which captures all microbial agents, including non-cultivable species.

Study design, participants, and setting
This was a secondary analysis of microbiological samples taken from participants in a randomised placebo-controlled trial of a daily oral probiotic combination [12e16]. The trial was approved by an NHS research ethics committee (Wales REC3. Ref: 15/WA/0306).
Participants were those living in a residential, nursing or dualregistered care home in England or Wales, aged 65 years or older, and not immunocompromised, or taking on-going regular probiotics who provided a stool sample at study entry (i.e. prior to randomisation). Prior to study participation, written informed consent was obtained from those participants with capacity to do so. For those who lacked capacity to provide consent, a consultee (either a family member or friend) completed a consultee declaration for participation on their behalf.

Procedures
Key data collected for the study included sex, age, weight, length of time resident in the care home, ethnicity, capacity status, clinical frailty [17], antimicrobial use in the previous 4 weeks, and use of proton pump inhibitors, laxatives, or vitamin D (included as there was interest in the original trial with the correlation between vitamin D and probiotic consumption) in the previous 4 weeks. A stool sample was collected from each participant using a clinically clean bedpan or receiver. Specimens were scooped in order to fill approximately one-third of the container (usually greater than 10 ml) and then placed in "Specisafe" packaging (a rigid plastic container to withstand postal damage) and posted via Royal Mail next day delivery, using compliant mailing envelopes (UN3373), to the Specialist Antimicrobial Chemotherapy Unit, Public Health Wales Microbiology, Cardiff, Wales. For transport, no recommendations were given regarding temperature ranges.

Microbiological analysis
Cultivable clinically relevant bacteria only were analysed. A 10 mL loop of stool sample was dissolved in 3 mL sterile saline and vortexed to mix. 50 mL was inoculated onto Columbia Blood Agar, plus a range of the Brilliance chromogenic agars to detect UTI pathogens (PO1110A), isolates harbouring antimicrobial resistance; ES B-Lactamase (PO0530A), Carbapenemase Resistant Enterobacterales (PO1226A) and Vancomycin Resistance Enterococci (PO1175A, Oxoid Ltd, Thermo Fisher, UK) using Spiral plater (WASP, Don Whitley, Bingley, Yorkshire UK). Plates were incubated in air at 35 C ± 1 C for 18e20 hours, then colony counts of all species calculated on each plate.
Cultures were then identified using a Bruker BioTyper Matrix Assisted Laser Desorption Ionisation e Time of Flight mass spectrometer.
Susceptibility testing by agar dilution [18] was performed on the predominant Gram-negative bacteria to a range of antimicrobials: amoxicillin, co-amoxiclav, gentamicin, trimethoprim, nitrofurantoin and ciprofloxacin. Minimum Inhibitory Concentrations were interpreted using the EUropean Committee on Antimicrobial Susceptibility Testing guidelines, version 2021 [19]. Antimicrobial resistance of isolates growing on the ESBL and Carbapenemase Resistant Enterobacterales media were confirmed using ROSCO Phenotyping kits (Rosco Diagnostica) to detect ESBL and ampC blactamases plus genotypic confirmation by in-house Real-Time Polymerase Chain Reaction assays for TEM, SHV, CTX-M genes. Minimum Inhibitory Concentrations for potential Vancomycin Resistance Enterococci were confirmed using vancomycin and teicoplanin gradient strips (Etest, BioMerieux, France) and an inhouse Real-Time Polymerase Chain Reaction assay for vanA/vanB genes.
Stool samples were also inoculated onto Fastidious Anaerobe Agar (FAA, Oxoid Ltd) and incubated in anaerobic conditions for 20e24 hours for Clostridioides difficile isolation. Any growth was identified using the Matrix Assisted Laser Desorption Ionisation e Time of Flight.

Outcomes
Outcomes included the total number of bacterial isolates cultured from stool samples, the frequency of different organisms cultured from stool samples, the number of different organisms per stool sample, the frequency of C. difficile, the susceptibility of organisms to antimicrobials, and the prevalence of antimicrobial resistance. Dietary intake was not standardised as part of the trial, and participants continued to follow their usual dietary habits. Care homes are expected to follow regulatory standards which highlight the importance of providing varied and nutritious meals, but meal provision will vary between care homes, and residents may have a less varied diet than community-dwelling individuals.

Statistical methods
Outcomes are presented descriptively, with frequencies provided with percentages, means with standard deviations, and medians with interquartile ranges as appropriate.
Care home variation was primarily explored descriptively, but the median odds ratio was estimated to assess the risk of having AMR Enterobacterales in a randomly chosen individual if they moved to a care home with a higher risk [20].

Results
The PRINCESS trial enrolled 310 participants from 23 care homes in England and Wales between December 2016 and May 2018. Out of these, 159 provided stool samples before randomisation (51% of all enrolled participants). Those who provided stool samples had a mean age of 85 years, with one-third being male, and 62% (n ¼ 99) lacked the capacity to consent to the study (Table 1).
One stool sample leaked in transit and was therefore not subject to analysis. In total, 402 clinically relevant bacterial pathogens were identified from 158 stool samples. The median total colony count was 9 Â 10 6 CFU/mL (IQR: 1 Â 10 6 to 3.9 Â 10 7 CFU/mL). Twentynine different species were identified (Fig. 1 The total number of different species cultured in any one sample ranged from one (14/158 samples, 9% of samples) to six (1/158 samples, 1% of samples), with a median of two different species cultured from a sample (IQR: 2 to 3 species per sample).
Care homes varied in the extent to which their recruited participants provided stool samples (from 0 to 85% of recruited participants) and the mean number of organisms per sample (from 1 to 3.8 organisms per sample). While the most commonly cultured species was E. coli across all care homes, the other most frequently cultured species varied (Table S1).
Four percent of samples, all of which were from participants from different care homes, grew C. difficile (toxin positive) on culture (6/158).
Enterobacterales isolates were cultured from 157 samples (99%), and resistance to at least one of the tested antimicrobials was found in 119 of these (76%).
The median odds ratio between care homes for the presence of AMR Enterobacterales was 2.96 (95% credible interval: 2.09 to 4.09), indicating substantial variability between care homes.
Overall, 39/157 (25%) participants had taken at least one antibiotic during the 4-weeks prior to study entry. While the number of participants taking trimethoprim in the 4-weeks prior to study entry was small (8/157, 5%), seven of these participants subsequently provided a stool sample containing trimethoprim-resistant Enterobacterales.
Three stool samples were confirmed positive for vancomycin resistant Enterococci, conferred by a vanA gene. Furthermore, 68 (43%) samples were found to contain bacteria that grew on ESBLselective media, suggestive of resistance to third-generation cephalosporins. Of these, 46% (31/68) were confirmed to harbour an ESBL (and thus in 20%, or 31/158 samples was there ESBL carriage) and the remainder (32 P. aeruginosa,twoCitrobacter freundii, and three Morganella morgannii) harboured a hyper-expressing native AmpC enzyme conferring third-generation cephalosporin

Discussion
This study of cultivable gut bacteria found that care home residents across the UK have a varied diversity of clinically relevant bacterial species. Stool samples from care home residents contained between one and six clinically relevant pathogens, some with high levels of resistance to antimicrobials used to treat common infections. The most common species found in the stools are also the most prevalent cause of blood stream infections [21]. A significant level of antimicrobial resistance was found, with the resistance rates similar to those nationally reported found in blood stream infection E. coli [21]. In this study, third-generation cephalosporin resistance in Enterobacterales was significantly higher than for those isolates causing bloodstream infections and UTI in UK population-based studies [21,22].
Unrecognised and untreated UTIs are often a causative factor in the development of systemic infections, and in the elderly, delayed prescriptions for UTIs have been associated with increased septicaemia [23]. E. coli, Enterococcus faecalis, K. pneumoniae and Proteus mirabilis are the most common pathogens associated with UTIs and were the most prevalent found in this study [24]. When considering antimicrobial agents commonly used to treat UTIs, resistance rates were similar for those found in both community and inpatient Welsh isolates for trimethoprim and ciprofloxacin [22,25]. Co-amoxiclav resistance rates in the E. coli from stools were higher than for Welsh UTI isolates. For the first line UTI therapy nitrofurantoin, the resistance rate in our study was zero compared with 3.1% and 3% found in the Welsh UTI community and in-patient isolates. These differences may reflect the settings in which these studies were carried out, the age ranges considered in the Welsh surveillance work, and the longer time frame over which the surveillance work was carried out. For nitrofurantoin, resistance rates are naturally low amongst urinary isolates in general, but the lack of resistance in our data set could be influenced by the low levels of nitrofurantoin available in the gut.
We found substantial variation among care homes regarding the presence of AMR Enterobacterales. This implies that in some care homes, the level of AMR Enterobacterales is high, consistent with previous work suggesting that the enclosed environments of care homes may act as a reservoir for the transmission of AMR infections [4,26].
The study has several limitations. Stool samples were provided by just over half of included residents. There may be factors which may indicate that these residents are a selective group. Identified barriers to the collection of stool samples relate to dignity and practical issues [27]. Future research aiming to study the gut microbiology of care home residents should consider the representativeness of those sampled and could consider different sampling strategies or oversampling in some circumstances.
This study has demonstrated that care home residents harbour significant levels of culturable AMR organisms in their gut. Given their living environment, increased frailty, and infection frequency (in particular UTIs), this work emphasises the importance of both enhanced infection control practices to limit the transmission of infections and antimicrobial stewardship programmes, aligned to guidelines around local resistance patterns, to support the appropriate use of antimicrobials in this setting. Further work is needed to understand transmission patterns within care homes, as well as the microbiome health and diversity in this population, as these *Ordered by most frequently occurring species. Data labels were provided for species where the number of isolates was !10. Colour coded by species type (blue ¼ Enterobacterales; orange ¼ pseudomonas species; green ¼ Enterococci species; yellow ¼ streptococcus species; grey ¼ other species). may be important areas to target for interventions aiming to reduce AMR transmission.

Author contributions
DG drafted the paper along with RR and MW. The statistical analyses were conducted by DG. Early drafts were reviewed by PC, TMML, and CCB. All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. DG and CCB have directly accessed and verified the underlying data reported in the manuscript.

Transparency declaration
All authors declare that they have no conflict of interests. This project was funded by the EME programme, which is funded by the MRC and NIHR, with contributions from the Chief Scientist Offi-cedScotland, Health and Care Research Wales, and the Health and Social Care Research and Development Division (part of the Public Health Agency) Northern Ireland (grant 13/95/10, Probiotics to Reduce Infections in Care Home Residents [PRINCESS]). The funding and sponsoring organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. In addition, they had no right to veto publication or to control the decision regarding which journal the paper was submitted.